Application of Neutrosophic Case-Based Reasoning and Neutrosophic Best-Worst Method for Product Cost Estimation

Abstract

The problem of product cost estimation is one of the crucial issues in contemporary manufacturing systems when new products arrive as customer orders. Product cost estimation determines the success of manufacturers. An overestimation causes them to lose sales and competitiveness and an underestimation results in a financial crisis. This should be done at the early of production to reduce the potential costs that can be incurred in the production, distribution, consumption, and disposal of products. In previous studies, this problem was addressed using different mathematical, heuristics, multiple-criteria decision-making (MCDM), and artificial intelligence (AI) methods. These methods have their advantages and disadvantages as stated by several studies. Referring to the previous studies, the integration of neutrosophic case-based reasoning (N-CBR) and neutrosophic best-worst method (N-BWM) was not applied to solve the problem of product cost estimation. This study aims to develop a decision support system (DSS) by integrating the neutrosophic versions of CBR and BWM to solve the problem of product cost estimation at the early production stage. This implies that the proposed system contributes additional knowledge to the current literature in product cost estimation and decision-making. This is because, nowadays, neutrosophic set theory (NST) is getting more attention to represent the knowledge of experts in MCDM. In this study, product orders were treated as multiple-attributed cases incorporating neutrosophic-based verbal terms, and numeric and categorical cost drivers as case attributes. In addition, this study applied an object-oriented programming (OOP) approach to represent part-order arrivals as cases with multiple attributes. Optimal weights of case attributes were determined using a group-based N-BWM. Neutrosophic-based verbal terms of cost drivers and neutrosophic BWM terms were converted into equivalent single-valued trapezoidal neutrosophic numbers (SVTNNs). From managerial implication, the proposed system can be applied to estimate the cost of new product orders at the early production stage in real manufacturing environments by integrating the proposed methodological approaches. In this study, a numerical example was illustrated in a simulated machining environment to test the soundness of the proposed system.